Insights

The e-commerce paradigm is entering its third inflection point. From consumers searching and paying themselves, we are moving toward agentic commerce — in which AI agents discover, compare, purchase, and pay on the consumer's behalf.

Samjong KPMG Economic Research Institute's Issue Monitor (February 2026) analyzes the e-commerce paradigm's evolution from search-driven to recommendation-driven to agentic commerce, and outlines Korean companies' response strategies. In agentic commerce, AI agents emerge as the purchasing actor, handling the full flow from product discovery to automated payment execution through autonomous reasoning and execution.

The Big Tech standards war: a protocol race for ecosystem hegemony. The most important variable in agentic commerce's spread is which protocol becomes the global standard. OpenAI proposes the execution-centric ACP (Actions in ChatGPT Protocol), through which agents discover products and complete payments. Google pursues end-to-end commerce-infrastructure control through AP2 (payment authentication and security) and UCP (a common language for the full commerce flow). Amazon aims to absorb the AI interface within its ecosystem via Rufus, while Perplexity seeks influence over the consumer's choice and decision moments. Just as HTML defined the web ecosystem, the first mover in the agentic-commerce protocol standard will hold ecosystem hegemony.

Korea: structural strengths in simple payments and digital infrastructure. Korea is slower than global firms in implementing agentic commerce, but holds structural strengths. A highly developed digital infrastructure and a universally adopted simple-payment ecosystem provide a favorable technology foundation for AI agents' autonomous transaction execution. A vast pool of consumer behavior data, high technology adoption, and dense platform environments connecting commerce with finance and messaging (Kakao, Naver) also create favorable conditions for agentic-commerce experimentation.

What companies must prepare: “agent-friendliness” as a new competitive standard. In the agentic-commerce era, competitiveness depends not only on being understood and chosen by consumers, but also on being understood and chosen by AI agents — i.e., “agent-friendliness.” Real-time inventory, pricing, and shipping data must be provided in standardized API formats, and the shift from exposure-centric marketing to AI-optimized operations is required. This is not a short-term wholesale transition but a process of building structural flexibility to prepare for change.

What this means for M&A. The rise of agentic commerce may trigger three deal flows. First, commerce-tech companies (payment, logistics, product-data platforms) with agent-commerce protocol compatibility and API-integration capabilities will emerge as strategic bolt-on targets. Second, AI solution companies with agent-based personalization, recommendation, and discovery capabilities may become acquisition targets for traditional commerce players in retail, finance, and home shopping. Third, e-commerce platforms left behind in the standards race may launch acquisitions of AI startups to internalize capabilities, or pursue strategic M&A to integrate into Big Tech ecosystems.